Legacy ImageMagick Discussions Archive

I think it happened in the last week (before yesterday, I'm pretty sure). And it still happens with this morning's svn (version 9629).
For example, does not show any verbose output.
Actually, it looks like -defining anything having to do with filters does not have an impact. For example,gives the exact same output as if I skip the -define.
-----
Hopefully I did not break things myself with a funky install, but I figured I'd put it out there.

Anthony this is a side effect of your latest changes, if we restore the option setting code in image.c/SyncImageSettings(), the define option works, however, with your new code:

• image->image_info = CloneImageInfo(image_info);

define settings are failing, i.e. resize.c/AcquireResizeFilter()/GetImageArtifact(image,"filter:filter") returns NULL. Note, without the CloneImageInfo() call, we get memory problems. Your original code generated a compile warning and fault:

• image->image_info =image_info;

Okay. The -define needs to be applied to each image. While 'set' sometimes needs to do this, other times it does not.

This is part of the seperation of a 'global' setting and per-image 'defines' (whcih can override a globle for a specific image).

One of the other changes I need to make is when a globel 'setting' is set, per-image defines of the same name are unset.

Actually this should work...
magick xc: -set filter:verbose 1 -distort Resize 200% null:

When the 'artifact lookup' fails. it is supposed to try looking up from the global settings (options in image_info). Obviously that is not happening, perhaps because the image->image_info setting is not getting set?

UPDATE: This should all be resolved. I discovered the fault in the last round of updates.

PS; that last example should have been


Another aspect is that if you undefine a option, you also need to undefine all per-image artifacts. This has also been fixed in IMv7.